한국판 DIAL-3의 민감도와 명확도, 동의지표

Abstract

The purpose of developmental screening is to identify children in need of further evaluation or guidance procedure. The most useful way to evaluate a screening instrument for developmental or primary care settings is to determine the test's accuracy in classification decisions. Overreferrals might lead to unnecessary worry among parents, and underreferrals might deprive children who need early intervention. The accuracy of the validity of a developmental screening instrument can be assessed in several ways. Sensitivity and specificity are suitable and for use in assessing the validity of developmental screening tests. Sensitivity refers to a test's accuracy in identifying all developmentally delayed cases (referring children who need special services). Specificity refers to a test's accuracy in identifying all of the non-developmentally delayed cases (not referring children who do not need special services). In general, sensitivity levels of 70% or more and specificity levels of 70-80% are acceptable. Tests that are highly sensitive would have fewer underreferrals (false negatives). Tests that are highly specific would have fewer overreferrals (false positive). The purpose of this study was to evaluate the validity of the Korean Developmental Indicators of Assessment for Learning 3 (K-DIAL 3) as a developmental screening instrument by examining the sensitivity, specificity, and agreement index as well as overrefferral and underreferral rates. This study also investigated the K-DIAL3 identification rate of children who need special services among participating children. A total of 109 children ages between 3-5 were sampled for this research. Two types of IQ tests were administered to the children after the K-DIAL 3 test. Sensitivity and specificity results were obtained from the use of cut-off scores. The results indicated that the K-DIAL 3 correctly identified 50-100% of the children who might need special services, demonstrating high rates of sensitivity. The optimal sensitivity achieved with the IQ 70 of K-ABC and all cut offs of K-DIAL 3, demonstrating 100% accuracy of identifying the children. The results also indicated that the K-DIAL3 identified correctly more than 93% of the children who may not need special services, regardless of the cut-off scores, demonstrating high rates of specificity. The overall agreement index of the K-DIAL 3 was .93-.95. It was found that the overreferral rates of K-DIAL 3 were 5-6% and underreferral rates were 0-3%.